Backstepping Controller with Neural Networks for MPPT in Photovoltaic Systems

This paper define the designing of a Backstepping Controller integrated with solar cell based photovoltaic (PV) and battery system. The combination of solar cell and batteries is installed to a single grid which will satisfy the need of the DC microgrid. Backstepping controller with neural network is designed for boost converter which is coupled with solar panel to extract maximum constant power and battery coupled with bi-directional converter to provide constant voltage to the grid. The responses obtained using backstepping controller with neural networks that is, steady state and transient state responses has proved to be better than the conventional PID controller. Tuning of the control laws in correspondence with Lyapunov stability analysis in order to achieve asymptomatically stable system is obtained. Further, using Hermite polynomial with five input variables to give the exact and best approximation solution of solar current with minimal errors. The robustness of the proposed controller is ensured by comparing the Backstepping Neural Network controller with the conventional PID controller. For the wide range of dissimilarities in the irradiance and grid voltage from the simulation results it is noticeable that the proposed controller shows the emphasized transient response with less settling time and overshoot and steady state response with minimum error and high efficiency.

Перевод пока недоступен